public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Adhemerval Zanella <adhemerval.zanella@linaro.org>
To: libc-alpha@sourceware.org
Subject: Re: add "workload" traces for sin/exp/pow in binary32/64/128
Date: Wed, 5 Aug 2020 16:33:19 -0300	[thread overview]
Message-ID: <33759faf-f40b-5c0b-c251-962d6d09a441@linaro.org> (raw)
In-Reply-To: <mw3651fhua.fsf@tomate.loria.fr>



On 05/08/2020 08:33, Paul Zimmermann wrote:
>        Hi,
> 
> this patch adds workload traces for sin/exp/pow in binary32/64/128.
> In addition, it adds traces for sin in the largest binade for each
> format, this enables to measure cases where a huge argument reduction
> is needed. Here are the corresponding results of "make bench" on my
> i5-4590 workstation (x/y means that reciprocal throughput is x clock
> cycles, latency is y clock cycles):
> 
>     function   binary32   binary64   binary128
>     sin        19/50      40/76      2444/2555
>     sinhuge    29/63      223/257    2619/2714
>     exp        16/39      21/44      2784/2869
>     pow        31/68      60/98      7740/7828
> 
> Each trace is made of 1000 random numbers, generated with SageMath.

For ex-inputs did you extend the expf-inputs input or the values were
generated randomly? If the later I think the benchtest name should
reflect it.

Also, for sinf-inputs the workload 'workload-spec2017.wrf' name be
misleading.

  reply	other threads:[~2020-08-05 19:33 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-05 11:33 Paul Zimmermann
2020-08-05 19:33 ` Adhemerval Zanella [this message]
2020-08-06  7:34   ` Paul Zimmermann
2020-08-25 16:05     ` Adhemerval Zanella
2020-08-26  8:06       ` Paul Zimmermann
2020-08-26 13:46         ` Adhemerval Zanella
2020-09-10 18:51           ` Adhemerval Zanella

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=33759faf-f40b-5c0b-c251-962d6d09a441@linaro.org \
    --to=adhemerval.zanella@linaro.org \
    --cc=libc-alpha@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).